The ns-1000's have a hugely wide baffle, is this a big factor in why there imaging is quite poor? annoying for a speaker that does everything else so well.
I have been thinking about how much the imaging might be improve if the drivers were in a narrow baffle for less diffraction. The 1000's image is like a wall of sound, i cant help thinking these speakers do not image well because the wide baffle.
I got reading about diffraction, here is a discribtion i have pasted it seems to say wide baffle = bad
The diffraction problem usually has more negative sonic consequences than the reflection problem. Sound produced by the transducers is diffracted by any sharp edges or discontinuities, causing additional phantom sources to appear in an improperly radiated sound field. An ideal sound field is created by a single point source. Diffraction produces a confusing superposition of phantom sources in addition to the primary point source, thereby creating multiple point sources that result in a less than ideal sound field.
Both effects will produce irregularities in the amplitude and phase response of the loudspeaker. The diffraction effects, in particular, tend to compromise the sound field imaging capability of the loudspeaker.
The vandersteen 5a speaker with it's amazing cabinet got me thinking. Might be a project i would try if it did improve the imaging.
I have been thinking about how much the imaging might be improve if the drivers were in a narrow baffle for less diffraction. The 1000's image is like a wall of sound, i cant help thinking these speakers do not image well because the wide baffle.
I got reading about diffraction, here is a discribtion i have pasted it seems to say wide baffle = bad
The diffraction problem usually has more negative sonic consequences than the reflection problem. Sound produced by the transducers is diffracted by any sharp edges or discontinuities, causing additional phantom sources to appear in an improperly radiated sound field. An ideal sound field is created by a single point source. Diffraction produces a confusing superposition of phantom sources in addition to the primary point source, thereby creating multiple point sources that result in a less than ideal sound field.
Both effects will produce irregularities in the amplitude and phase response of the loudspeaker. The diffraction effects, in particular, tend to compromise the sound field imaging capability of the loudspeaker.
The vandersteen 5a speaker with it's amazing cabinet got me thinking. Might be a project i would try if it did improve the imaging.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
I also owned a pair of NS1000's many years ago. Yes, imaging is the weakest point in an otherwise brilliant speaker.
In my quest for improved imaging I did conduct various experiments. I think you can get a fair idea of what the Vandersteen concept will achieve without a big effort and expenditure.
Here's what I did:
Turned the cabinet upside down to bring the woofer closer to the top - still stand mounted.
Removed the midrange and tweeter drivers and wired them for mounting on top of the inverted cab.
Sealed the holes left by the drivers temporarily.
Made temporary fixtures to keep the drivers in position on top of the inverted cabinet, much like the Vandersteen you referenced.The baffle widths were therefore only determined by the driver face plates.
I tried to balance the speaker by using just the level controls. This will of course not give you an ideal linear response, but it is still useful to get an idea of the imaging that such a layout may provide.
I was pleasantly surprised by the improvement in imaging. The sound seemed to detach from the speakers and hung in space. I then added temporary roundovers to the driver face plates using modeling clay. This brought about further improvement, although not of the magnitude the the narrower baffles (face plates only) did.
I did detect some frequency response irregularities, but not as much as I expected. But in the end one will have to redesign the crossover. I never got around to doing that.
In my quest for improved imaging I did conduct various experiments. I think you can get a fair idea of what the Vandersteen concept will achieve without a big effort and expenditure.
Here's what I did:
Turned the cabinet upside down to bring the woofer closer to the top - still stand mounted.
Removed the midrange and tweeter drivers and wired them for mounting on top of the inverted cab.
Sealed the holes left by the drivers temporarily.
Made temporary fixtures to keep the drivers in position on top of the inverted cabinet, much like the Vandersteen you referenced.The baffle widths were therefore only determined by the driver face plates.
I tried to balance the speaker by using just the level controls. This will of course not give you an ideal linear response, but it is still useful to get an idea of the imaging that such a layout may provide.
I was pleasantly surprised by the improvement in imaging. The sound seemed to detach from the speakers and hung in space. I then added temporary roundovers to the driver face plates using modeling clay. This brought about further improvement, although not of the magnitude the the narrower baffles (face plates only) did.
I did detect some frequency response irregularities, but not as much as I expected. But in the end one will have to redesign the crossover. I never got around to doing that.
That vandersteen illustration serves its marketing purpose, but it doesn't accurately illustrate real-world acoustic behavior.
A better way to to visualize how sound waves propagate from drivers in baffles is to use Axidriver or Ripple (specify plane wave in Ripple and drag an edge to widen or narrow the transducer)
Baffle width must always be considered in relation to driver diameter and operating frequency. For example, an 8-inch woofer in a 2-foot baffle at 1kHz will radiate a pattern similar to a 4" midrange in a 1-foot baffle at 2kHz and a 2" mid/tweeter in a 6" baffle at 4kHz (notice locked relationship of driver and baffle size to wavelength).
All this is to say that, depending on driver diameter and operating frequency, a visually slim baffle may actually be "acoustically" wide, or a visually wide baffle may be "acoustically" narrow.
In the same way, a round-over on baffle edges is only effective in preventing diffraction if its radius is significant in relation to the radiated wavelength.
I'm no expert in the science behind the psycho-acoustic phenomenon called "imaging," but the 1987 Archimedes Study of loudspeaker/room interactions and subjective fidelity showed that power response should be flat and smooth, front-wall reflections were most to be avoided, followed by strong floor & ceiling reflections. (Interestingly, sidewall reflections were not found to be bad. David Moulton, for one, in fact claims they're highly beneficial.)
These findings could be interpreted to recommend wide-dispersion drivers on wide baffles (to prevent rearward radiation) with significant edge treatment to minimize diffraction.
Narrow-baffle loudspeakers often do wide-dispersion well, but beware their tendency to radiate backwards. Done wrong, they can also exhibit extreme transitions from 4pi radiation to <2pi radiation, causing wildly varying power response.
A better way to to visualize how sound waves propagate from drivers in baffles is to use Axidriver or Ripple (specify plane wave in Ripple and drag an edge to widen or narrow the transducer)
Baffle width must always be considered in relation to driver diameter and operating frequency. For example, an 8-inch woofer in a 2-foot baffle at 1kHz will radiate a pattern similar to a 4" midrange in a 1-foot baffle at 2kHz and a 2" mid/tweeter in a 6" baffle at 4kHz (notice locked relationship of driver and baffle size to wavelength).
All this is to say that, depending on driver diameter and operating frequency, a visually slim baffle may actually be "acoustically" wide, or a visually wide baffle may be "acoustically" narrow.
In the same way, a round-over on baffle edges is only effective in preventing diffraction if its radius is significant in relation to the radiated wavelength.
I'm no expert in the science behind the psycho-acoustic phenomenon called "imaging," but the 1987 Archimedes Study of loudspeaker/room interactions and subjective fidelity showed that power response should be flat and smooth, front-wall reflections were most to be avoided, followed by strong floor & ceiling reflections. (Interestingly, sidewall reflections were not found to be bad. David Moulton, for one, in fact claims they're highly beneficial.)
These findings could be interpreted to recommend wide-dispersion drivers on wide baffles (to prevent rearward radiation) with significant edge treatment to minimize diffraction.
Narrow-baffle loudspeakers often do wide-dispersion well, but beware their tendency to radiate backwards. Done wrong, they can also exhibit extreme transitions from 4pi radiation to <2pi radiation, causing wildly varying power response.
Last edited:
I think there is something to the narrow versus wide baffle in terms of imaging quality. It has been my observation (subjective as it is) that narrower baffles do image better and that as a baffle gets wider (at least in the midrange and tweeter area), it eventually becomes easier to audibly locate the speaker in the room (point to its location with you eyes closed).
These two observations have led me to the design I'm working on now. It uses somewhat progressively narrower baffles as you go from the low frequency driver, to the woofers, midrange and tweeter. It is a 4way (SWWMT) with a 10" sub and active crossover at around 100hz on an 11" wide baffle (no chamfer). The WWMT portion of the speaker will be passive. Two 7" woofers are on an 11" wide baffle with 1" chamfers, crossing over to a 2" mid dome at about 1000hz on a 5.5" baffle with 1" chamfers, that crosses to a 3/4" dome tweeter at around 3500hz also on the same 5.5" baffle with 1" chamfers.
The depth of the boxes are designed to reduce the amount of the sound does not wrap around the baffle and the sides and radiate back toward the front wall. This should do away with front wall reflections about 600hz or so and up, allowing for the speaker to be placed closer to the wall than typical, yet provide excellent soundstage width and imaging (hopefully). Also, the intent of the narrower baffle is to try and push the diffraction dips up into the drivers stop band, where possible, in an attempt to reduce the its impact.
This design is being done with drivers I already have on hand. The sub driver is a 10" Dayton Reference Series RSS265HF in a tapered transmission line with an F3 of around 25hz or so and an F6 of around 19hz. With room and boundary gain, it should be pretty flat to 20hz or so. The woofers are Dayton RS180's (in a sealed box). The midrange is the Morel MDM55 and the tweeter is the Seas 22TAFG. I'm calling them "The Blades" due their resemblance to a utility knife.
Below is what the design concept looks like:
These two observations have led me to the design I'm working on now. It uses somewhat progressively narrower baffles as you go from the low frequency driver, to the woofers, midrange and tweeter. It is a 4way (SWWMT) with a 10" sub and active crossover at around 100hz on an 11" wide baffle (no chamfer). The WWMT portion of the speaker will be passive. Two 7" woofers are on an 11" wide baffle with 1" chamfers, crossing over to a 2" mid dome at about 1000hz on a 5.5" baffle with 1" chamfers, that crosses to a 3/4" dome tweeter at around 3500hz also on the same 5.5" baffle with 1" chamfers.
The depth of the boxes are designed to reduce the amount of the sound does not wrap around the baffle and the sides and radiate back toward the front wall. This should do away with front wall reflections about 600hz or so and up, allowing for the speaker to be placed closer to the wall than typical, yet provide excellent soundstage width and imaging (hopefully). Also, the intent of the narrower baffle is to try and push the diffraction dips up into the drivers stop band, where possible, in an attempt to reduce the its impact.
This design is being done with drivers I already have on hand. The sub driver is a 10" Dayton Reference Series RSS265HF in a tapered transmission line with an F3 of around 25hz or so and an F6 of around 19hz. With room and boundary gain, it should be pretty flat to 20hz or so. The woofers are Dayton RS180's (in a sealed box). The midrange is the Morel MDM55 and the tweeter is the Seas 22TAFG. I'm calling them "The Blades" due their resemblance to a utility knife.
Below is what the design concept looks like:
Attractive looking speakers, Dan.
If I may make a suggestion, it would be smart to model your design in LEAP or another ap that can predict each driver's power response around the crossover points. At 1kHz, transitioning from an 11" baffle loading your midbass at roughly 2pi to a 5.5" baffle loading the midrange at roughly 4pi means you could have significant ripple in the power response. Definitely something worth checking, IMO.
Also, at 1kHz, the 1" chamfer on your midbass baffle would be purely cosmetic (it's only about a tenth of a wavelength), but still worth keeping because it does look good.
(PS: what part of IN? I'm only a few miles north of the border.)
If I may make a suggestion, it would be smart to model your design in LEAP or another ap that can predict each driver's power response around the crossover points. At 1kHz, transitioning from an 11" baffle loading your midbass at roughly 2pi to a 5.5" baffle loading the midrange at roughly 4pi means you could have significant ripple in the power response. Definitely something worth checking, IMO.
Also, at 1kHz, the 1" chamfer on your midbass baffle would be purely cosmetic (it's only about a tenth of a wavelength), but still worth keeping because it does look good.
(PS: what part of IN? I'm only a few miles north of the border.)
OK... like so many... have spent a lot of time with the 1000... wonderful piece of engineering... did my share of experiments... discussions, measurements and reached the following conclutions...
1) The crossover was designed to sum flat on axis... typical of a near field monitor and thought to be THE way to go when the speaker was designed.
2) But... one of the penalties is a less that perfect power response. This is what really hurts the imaging ability of the speaker. The sharp box edges sure don't help... but the offset of the drivers helps to offset some of the edge diffraction problems.
3) Another concession to a near field monitor design is the lack of bass step correction... with the speaker intended to be close to a rear wall in a small control room environment it was thought the speakers location and the cabin gain typically found in a small room would fill out the bass end of things... and... the engineers placed big priorities on dynamic potential as well as reasonable efficiency for the box size.
A lot of this was related to me by Yamaha engineers I spoke with at CES in the 70's and it makes good sense.
I have re-worked the crossover for better power response and added about 3 db of bass step and this greatly improved the tonal balance and imaging. I used the speaker in a typical residential environment.. on 24" stands about 18" from the rear room walls. This was over 20 years ago... but I do have some great memories from listening to the re-worked version. Much better suited to residential HI-FI listening chores. If you want to hear what the 1000 was intended to do... put it in a small bed room at 3-4 ft. above the floor up against the wall and listen to it near field... WOW!!
1) The crossover was designed to sum flat on axis... typical of a near field monitor and thought to be THE way to go when the speaker was designed.
2) But... one of the penalties is a less that perfect power response. This is what really hurts the imaging ability of the speaker. The sharp box edges sure don't help... but the offset of the drivers helps to offset some of the edge diffraction problems.
3) Another concession to a near field monitor design is the lack of bass step correction... with the speaker intended to be close to a rear wall in a small control room environment it was thought the speakers location and the cabin gain typically found in a small room would fill out the bass end of things... and... the engineers placed big priorities on dynamic potential as well as reasonable efficiency for the box size.
A lot of this was related to me by Yamaha engineers I spoke with at CES in the 70's and it makes good sense.
I have re-worked the crossover for better power response and added about 3 db of bass step and this greatly improved the tonal balance and imaging. I used the speaker in a typical residential environment.. on 24" stands about 18" from the rear room walls. This was over 20 years ago... but I do have some great memories from listening to the re-worked version. Much better suited to residential HI-FI listening chores. If you want to hear what the 1000 was intended to do... put it in a small bed room at 3-4 ft. above the floor up against the wall and listen to it near field... WOW!!
Hi Bill,
I'm a SoundEasy user and this is all prior to any crossover modeling. I will be building a prototype to do actual measurements from. The top MT section will remain removable in case there are issues the require a design alteration. For example, I could step the baffle width further by addding additional width in the midrange area and step back again in the tweeter area. This is just the first steps in the process.
What I have modeled so far is baffle diffraction, also combined with room and boundary gain modeling and TL modeling using MJK's mathcad worksheets. The 1" chamfer on the 11" baffle does do make some difference in terms of reducing the diffraction peak and dip, thus flattening the response, but I doubt it would be audible, especially since the diffraction ripple is up around 1500hz, IIRC.
Actually, let me post a couple of the model results (see below), since they are somewhat germain to the discussion of baffle width. The first two are the bottom 7" woofer with the chamfer followed by the bottom woofer without chamfer. Below that are the top 7" woofer with the chamfer, followed by top woofer without chamfer. Below that are the midrange and tweeter diffraction models. I also have models for each driver that combine diffraction, room gain and boundary gain in the intended room and speaker position as well.
BTW, I'm in Bloomingon but grew up in Michigan and went to MSU!
Lower 7" wooofer
Upper 7" woofer
Midrange
Tweeter
Yes i can say for a fact that the NS1000M's will image vastly better with a narrower baffle & rounded edges, i happen to have 3 pairs of them. Ok, so only one pair is in use right now as the other 2 pairs are going to make new front speakers.
Here's a picture of the ones in use.
Basic rundown is that the 12" bass driver is now in a 4th order bandpass which cures the weak bass, these go down to just under 30Hz -3db with no equalisation. The 12" has been replaced by a 7" Audax bass that took me an age to find so it'd work well with the original crossover. An electronic crossover splits between the 12" which is active & the Yam crossover at 110Hz. So i need a stereo power amp per channel to drive them.
Enclosure is now 36mm thick MDF heavliy braced. I'm a happy chappie 😀 These will become rear speakers in a 5.1 setup, i'll not take these apart or sell them like i have other stuff.
It's definately woth giving them better enclosures if you have the inclination & can find the time.
Here's a picture of the ones in use.
Basic rundown is that the 12" bass driver is now in a 4th order bandpass which cures the weak bass, these go down to just under 30Hz -3db with no equalisation. The 12" has been replaced by a 7" Audax bass that took me an age to find so it'd work well with the original crossover. An electronic crossover splits between the 12" which is active & the Yam crossover at 110Hz. So i need a stereo power amp per channel to drive them.
Enclosure is now 36mm thick MDF heavliy braced. I'm a happy chappie 😀 These will become rear speakers in a 5.1 setup, i'll not take these apart or sell them like i have other stuff.
It's definately woth giving them better enclosures if you have the inclination & can find the time.
I have built narrow and wide baffles. Sure, narrow baffles have some appeal, maybe giving a more spatial impression. However they suffer from secondary emission due to diffraction arriving in a too short delay after the primary emission of the drivers and the acoustic image is less precise. Large baffles avoid this, that was the idea behind the Snell model A. I built two prototypes (ugly...) 2' wide with rounded edges, they provide nice sharp images. I got inspiration from a Kef hi-priced model and by the Stradivari Homage (or Elipsa) which can be seen here
:: SONUS FABER ::
It would be interesting to compare it with the Amati Anniversario (or the Cremona M) by the same manufacturer, which seems to have identical drivers but a radically opposite philisophy, with a narrow baffle.
:: SONUS FABER ::
It would be interesting to compare it with the Amati Anniversario (or the Cremona M) by the same manufacturer, which seems to have identical drivers but a radically opposite philisophy, with a narrow baffle.
Very unique design, how can you tell they image well if your using them in 5.1?
I'm not right now as i think i mentioned. At the present they are the front speakers in my hifi system, they get used every evening unless i happen to be out.
Front end is a Krell DT-10 transport, Monarchy Audio DAC feeding an Audio Research MP1 pre, Parasound HCA1206 (using 4 channels) to drive the speakers. The active crossover is homemade single ended class A design, i think i have a pretty good idea what they can do 😉 It's probably why i don't go out that often
I wouldn't even consider them NS1000's any more. They are a new design using some of the NS parts.
It would be. My experience with baffles is that I prefer wider, with a large chamfer or radius. The subjective reviews on these wide SF models and Troels Grevesen's comments on his WB designs are interesting.
Here's a picture of the ones in use
Very similar shape to a set of Fonken. They are image champs (not as much with stock drivers as treated ones). Big 3D soundstage, speakers are mostly unlocatable sonically.
dave
Dave, this speaker/room interaction issue is my current hobby horse. Can you link me to Toole's findings? (No AES membership here, sadly.)
A pair of AR LST II speakers I picked up at a yard sale got me curious. They were designed primarily for flat power response in an echo chamber. On-axis FR looks dreadful, yet, to my ears, they do so much right.
On my drawing board now is a speaker with a design goal of true 180-deg. horizontal dispersion up to ~5kHz or better, just to see how it sounds with a vertically-aligned array without the time-domain chaos of LST's horizontal array. Nothing else I'm aware of comes close to that spec, with the possible exception of some wacky ribbons. (I know, omnipoles do 360-deg., but I'm interested in eliminating front-wall reflections.)
Last edited:
Bill...
You have... of coarse... hit upon the key element of speaker design for residential applications... Power Response...
Wide baffles are better... all things equal... but WAF and other cosmetic issues prefere tall... slim baffle faces... So... much of what is done these days is geared to produce the best compromise a slim baffle requires. Too bad.. but that's the way it is.
You have... of coarse... hit upon the key element of speaker design for residential applications... Power Response...
Wide baffles are better... all things equal... but WAF and other cosmetic issues prefere tall... slim baffle faces... So... much of what is done these days is geared to produce the best compromise a slim baffle requires. Too bad.. but that's the way it is.
Amazon.com: Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms (9780240520094): Floyd Toole: Books
Sadly what is on the net is a bit lacking, i had read a lot of it and was expecting to be at loggerheads with what he had to say... turns out his position is not that far off and it reinforced, refined, and extended many of my feelings.
Sean Olive does touch on much of it.
dave
Turns out both are important, but neither by themselves is a good indicator.
dave
- Status
- Not open for further replies.